ModelDeepSeek V3.2-Reasoning / V3.2-Reasoning

DeepSeek V3.2-Reasoning

A reasoning-capable model version released by DeepSeek, showing relatively strong performance on structured generation tasks such as Text-to-SQL.

Timeline (last 90 days)

Oct 17

On the Spider dataset, schemagate tied with the same model (Claude Sonnet 4.5) at 93.0; the paper's best result on Spider came from DeepSeek V3.2-Reasoning at 95.9.

Unverified50%

All Facts (1)

Unverified

On the Spider dataset, schemagate tied with the same model (Claude Sonnet 4.5) at 93.0; the paper's best result on Spider came from DeepSeek V3.2-Reasoning at 95.9.

50%

Source Articles

Hiding Tables Won't Stop Text-to-SQL? schemagate Retriever's Real-World CounterOct 17